一种改进的时延受限低代价光组播路由算法

针对光网络时延受限光组播路由计算复杂度高的问题,提出一种基于改进的和声搜索求解时延受限光组播路由的算法。论文通过引入自适应的和声记忆保留率及微调概率提高和声算法搜寻到全局最优光组播树的能力;算法设计了一个能够保持备选光组播树多样性的动态适应度函数以扩大光组播路由的搜索范围;同时,算法在初始化和声库时采用精英保留策略简化和声迭代过程。仿真结果表明,提出的改进算法能够求得代价更低的光组播路由,且算法在较大光网络规模下具有较好的收敛稳定性。     

一种快速低代价延迟受限组播路由算法

为满足实时业务的QoS要求，在非延迟受限组播路由算法(Fast Low-cost Shortest Path Tree，FLSPT）的基础上添加了延迟约束，使得生成的组播树上，每条从源到目的地的路径都满足给定的延迟限制，同时保持了原算法计算复杂度低，代价性能优越的特点。仿真结果表明，本文算法的代价和时间性能均优于延迟受限最短路径(Delay-Constrained Shortest Path, DCSP)，且更适合用于目的节点分布集中的密集模式下。     

一种时延受限的组播路由算法

本文给出了时延约束组播路由问题的数学模型,提出了一种分布式、收敛快、支持动态组播的时延约束组播路由算法--DMPH(Delay-constrained Minimal-cost Path Heuristic),分析表明采用该算法具有良好的网络开销性能.     

时延受限的移动sink数据收集算法

在实时性要求比较高的应用中,时延要求限制了sink的移动速率与移动轨迹,sink的移动速率限制了节点与sink的通信时间,因此很难兼顾时延要求与数据收集效率。提出一种时延受限的移动sink数据收集算法MSDC,在低能耗缓存区内找到一条sink的最优移动轨迹,在有限的时间限制内利用sink的移动性来提升传感器网络的数据收集性能。仿真结果表明,与已有算法比较,该方案能够提高网络数据采集量,降低能耗,延长网络生命周期。     

DC-Gear: 延时约束型无线传感网占空比调控策略研究

Many Wireless Sensor Network (WSN) systems are deployed in unattended areas using non-rechargeable batteries. To enable sustainable operations, most WSN systems employ duty-cycling mechanisms, such as Low Power Listening (LPL). For reliable delivery of each packet with LPL, the sender has to transmit a preamble that is long enough to span over a complete sleep interval of the receiver. In this way, the sensor nodes avoid idle listening, however, at the cost of remarkably increased end-to-end delay of multi-hop packet transmissions. To address this issue, in this paper we propose a new duty-cycling mechanism called DC-Gear. DC-Gear exploits a“sleep less but save more” phenomenon, which means increasing the duty cycle in a timely and appropriate manner while minimizing the overall energy cost and satisfying the end-to-end delay constraint. We have implemented DC-Gear with TelosB motes and demonstrated its performance advantages through extensive experiments.     

A Polynomial Time Approximation Scheme for Minimum Cost Delay-Constrained Multicast Tree under a Steiner Topology

This paper is concerned with a restricted version of minimum cost delay-constrained multicast in a network where each link has a delay and a cost. Given a source vertex $s$ and $p$ destination vertices $t_1, t_2, \ldots, t_p$ together with $p$ corresponding nonnegative delay constraints $d_1, d_2, \ldots, d_p$, many QoS multicast problems seek a minimum cost multicast tree in which the delay along the unique $s$--$t_i$ path is no more than $d_i$ for $1 \le i \le p$. This problem is NP-hard even when the topology of the multicast tree is fixed. In this paper we show that every multicast tree has an underlying Steiner topology and that every minimum cost delay-constrained multicast tree corresponds to a minimum cost delay-constrained realization of a corresponding Steiner topology. We present a fully polynomial time approximation scheme for computing a minimum cost delay-constrained multicast tree under a Steiner topology. We also present computational results of a preliminary implementation to illustrate the effectiveness of our algorithm and discuss its applications.     

基于遗传算法的带时延约束组播路由算法

探讨了带时延约束组播路由优化算法,选用时延约束信息产生备选路径集并编码,给出了在该编码方式下使用不同进化阶段应用不同变异概率思想的改进遗传算法.仿真试验结果表明,该算法是可行有效的.     

一种改进的时延受限多播路由算法

针对DCMPH算法不能合理选择连接路径的问题,提出一种改进的满足时延限制的多播路由算法。该算法对不能用最小代价路径连接到多播树上的目的节点,求出其到多播树上所有节点的最小时延路径,再从中选出一条能满足时延限制的费用最小的路径,添加到多播树上。实验结果表明,与DCMPH算法相比,该算法构造多播树的代价更低。     

A Polynomial Time Approximation Scheme for Minimum Cost Delay-Constrained Multicast Tree under a Steiner Topology

This paper is concerned with a restricted version of minimum cost delay-constrained multicast in a network where each link has a delay and a cost. Given a source vertex $s$ and $p$ destination vertices $t_1, t_2, \ldots, t_p$ together with $p$ corresponding nonnegative delay constraints $d_1, d_2, \ldots, d_p$, many QoS multicast problems seek a minimum cost multicast tree in which the delay along the unique $s$--$t_i$ path is no more than $d_i$ for $1 \le i \le p$. This problem is NP-hard even when the topology of the multicast tree is fixed. In this paper we show that every multicast tree has an underlying Steiner topology and that every minimum cost delay-constrained multicast tree corresponds to a minimum cost delay-constrained realization of a corresponding Steiner topology. We present a fully polynomial time approximation scheme for computing a minimum cost delay-constrained multicast tree under a Steiner topology. We also present computational results of a preliminary implementation to illustrate the effectiveness of our algorithm and discuss its applications.     

基于拉格朗日松弛法的时延约束组播路由算法

通过对时延约束组播路由网络模型的分析,提出了一种基于拉格朗日松弛法的时延约束的低代价组播路由算法(LR-DLMR)。由于封闭图对原网络的多播不可达问题,该算法并没有构建原网络的封闭图,从而有效利用了链路中间节点信息。仿真实验结果表明本算法具有良好的稳定性,有较低的代价和时延。